| 摘要 |
<p><P>PROBLEM TO BE SOLVED: To provide a voiceprint authentication system utilizing vowel retrieval which can shorten the time required for collation. <P>SOLUTION: The voiceprint authentication system utilizing the vowel retrieval to perform authentication by pronounced speeches has a voiceprint data forming section which forms voiceprint data from the speeches, a text data forming section which forms text data from the speeches, a vowel extracting section which forms vowel key information by extracting the vowels included in the text data formed by the text data forming section, a voiceprint data base which prestores the voiceprint data and the voiceprint data while making both data correspond to each other, a retrieval section which retrieves the vowel data coinciding with the vowel key information formed by the vowel extracting section, and voiceprint collating section which performs authentication by collating the voiceprint data made to correspond to the vowel data coinciding with the vowel key information with the voiceprint data formed by the voiceprint data forming section on the basis of the retrieval results of the retrieval section. <P>COPYRIGHT: (C)2004,JPO</p> |
